湖北汽车工业学院学报2025，Vol.39Issue(4)：13-19,7.DOI:10.3969/j.issn.1008-5483.2025.04.003
面向混合型数据的邻域条件熵特征选择算法
Neighborhood Conditional Entropy Feature Selection Algorithm for Hybrid Data
摘要
Abstract
A hybrid neighborhood relation and an extended neighborhood rough set model were intro-duced.Based on this model,hybrid neighborhood information entropy,joint entropy,and conditional en-tropy were proposed,and the effectiveness of hybrid neighborhood conditional entropy as a feature eval-uation criterion for datasets was validated through theoretical analysis.Feature evaluation was per-formed using hybrid neighborhood conditional entropy,and a heuristic feature selection algorithm was designed based on a greedy strategy.Experimental results show that compared with the other four fea-ture selection algorithms,the proposed algorithm reduces the number of selected features by 8.2%,1.9%,8.1%,and 17.2%,improves the classification accuracy of feature subsets by 1.8%,1.7%,1.6%,and 5.3%,and significantly reduces feature selection time.关键词
混合型数据/特征选择/粗糙集/信息熵/条件熵Key words
hybrid data/feature selection/rough set/information entropy/conditional entropy分类
信息技术与安全科学引用本文复制引用
Zhang Hongyuan,Xie Jin..面向混合型数据的邻域条件熵特征选择算法[J].湖北汽车工业学院学报,2025,39(4):13-19,7.基金项目
安徽省教育厅自然科学重点项目(2024AH050900 （）
2023AH051450 （）
2023AH051452) （）
